MedWatch

Reduce the load on overwhelmed hospitals with a remote patient monitoring solution.

Comment

Inspiration

My inspiration was the fact that cloud technologies are the future of IT infrastructure. Currently many hospitals are overflowing with COVID-19 patients. Because of this, many patients with severe symptoms aren’t able to get adequate medical attention . A significant number of admitted COVID-19 patients may not need intensive in-person care but should still be monitored for health deterioration. What if these patients could instead be monitored remotely, at a fraction of the price?

What it does

Introducing MedWatch. Doctors send milder patients home with all necessary medical sensors to measure vitals and any other time series data points and either MedWatch agent device or smartphone app which sends sensor data to cloud-based MedWatch software application. If patient shows signs of deterioration, patient can be called in to hospital or nurse/emergency services can be sent to the patient’s home. Doctors and Nurses remotely monitor MedWatch patient dashboard. Time series data points from variety of medical sensors is archived in secure MedWatch cloud database which is HIPAA-compliant. Variety of sensor data is dynamically cleaned/curated into standardized format. Time series data can be easily analyzed by AI/ML algorithms as part of clinical studies, precision medicine research, and improving patient care.

How I built it

AWS cloud technologies. These include (AWS): IoT, API Gateway, Lambda, DynamoDB, Kinesis Firehose, and S3. AWS IoT receives simulated medical sensor data. This triggers a Lambda function which updates the record in DynamoDB table. A static website written in HTML and Javascript hosted in S3 performs an HTTP GET request to API Gateway which then triggers a Lambda function to retrieve latest records from DynamoDB table based (currently) on patient name (though something like Bed name or ward number could also be used if a hospital prefers). In addition to this, data from AWS IoT is sent to Kinesis Firehose and cleaned into various output data streams using SQL queries. These output data streams have S3 configured as a destination, so the cleaned medical sensor data is stored securely and durably in AWS S3 in separate folders, each folder corresponding to an output data stream.

Challenges I ran into

Configuration of API Gateway and IAM permissions.

Accomplishments that I'm proud of

Utilizing state-of-the-art cloud technologies on AWS.

What I learned

How to identify, describe, and address the key components of a successful Business Model.

What's next for MedWatch

AI/ML algorithms to send alerts tailored to each patient's condition and Doctor specifications to prevent alert fatigue. Video streaming of patient for on-demand Doctor consultation and additional dimension of video-based monitoring by AI/ML recognition algorithms for critical condition.

Built With

Share this project:

Updates